LoRA-FLUX-anime-style-v1

<details><summary>See ostris/ai-toolkit config</summary>

ostris/ai-toolkit commit hash: fa187b1

job: "extension"
config:
  name: "FAS"
  process:
    - type: "ui_trainer"
      training_folder: "/data/minpeter/github.com/flux/ai-toolkit/output"
      sqlite_db_path: "./aitk_db.db"
      device: "cuda"
      trigger_word: "anime style"
      performance_log_every: 10
      network:
        type: "lora"
        linear: 32
        linear_alpha: 32
        lokr_full_rank: true
        lokr_factor: -1
      save:
        dtype: "bf16"
        save_every: 250
        max_step_saves_to_keep: 4
        save_format: "diffusers"
        push_to_hub: false
      datasets:
        - folder_path: "/data/minpeter/github.com/flux/ai-toolkit/datasets/fas"
          mask_path: null
          mask_min_value: 0.1
          default_caption: "anime style, "
          caption_ext: "txt"
          caption_dropout_rate: 0.05
          cache_latents_to_disk: false
          is_reg: false
          network_weight: 1
          resolution:
            - 512
            - 768
            - 1024
            - 1280
            - 1536
            - 256
        - folder_path: "/data/minpeter/github.com/flux/ai-toolkit/datasets/f"
          mask_path: null
          mask_min_value: 0.1
          default_caption: ""
          caption_ext: "txt"
          caption_dropout_rate: 0.05
          cache_latents_to_disk: false
          is_reg: false
          network_weight: 1
          resolution:
            - 512
            - 768
            - 1024
            - 1280
            - 1536
            - 256
      train:
        batch_size: 1
        bypass_guidance_embedding: false
        steps: 2500
        gradient_accumulation: 1
        train_unet: true
        train_text_encoder: false
        gradient_checkpointing: true
        noise_scheduler: "flowmatch"
        optimizer: "adamw8bit"
        timestep_type: "sigmoid"
        content_or_style: "balanced"
        optimizer_params:
          weight_decay: 0.0001
        unload_text_encoder: false
        lr: 0.0001
        ema_config:
          use_ema: true
          ema_decay: 0.99
        dtype: "bf16"
        diff_output_preservation: false
        diff_output_preservation_multiplier: 1
        diff_output_preservation_class: "person"
      model:
        name_or_path: "black-forest-labs/FLUX.1-dev"
        quantize: true
        quantize_te: true
        arch: "flux"
        low_vram: false
      sample:
        sampler: "flowmatch"
        sample_every: 250
        width: 1024
        height: 1024
        prompts:
          - "anime style, woman with red hair, playing chess at the park, bomb going off in the background"
          - "anime style, a man showing off his cool new t shirt at the beach, a shark is jumping out of the water in the background"
          - "anime style, a bear building a log cabin in the snow covered mountains"
        neg: ""
        seed: 42
        walk_seed: true
        guidance_scale: 4
        sample_steps: 25
        num_frames: 1
        fps: 1
meta:
  name: "[name]"
  version: "1.0"

</details><br>

Fantasy Anime Style

A Flux LoRA trained on a local computer with ostris/ai-toolkit.

<Gallery />

Trigger words

You should use anime style to trigger the image generation.

Creating images using a diffusers

import torch
from diffusers import FluxPipeline

pipe = FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16).to("cuda")

from huggingface_hub import hf_hub_download

pipe.load_lora_weights(
    hf_hub_download("minpeter/FAS-v1", "adapter.safetensors"), adapter_name="fas-v1"
)

pipe.set_adapters(["fas-v1"], adapter_weights=[1])

prompt = "anime style, A vibrant fantasy city with sky-high spires and unique architectural buildings stacked upon each other, featuring bright and colorful tones with intricate details – anime style background."

image = pipe(
    prompt=prompt,
    guidance_scale=3.5,
    height=1024,
    width=1024,
    num_inference_steps=20,
).images[0]

image.save("s4.png")